Browsing by Author "Zaporozhets, Iryna"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Accurate nuclear quantum statistics on machine-learned classical effective potentials
    (AIP Publishing, 2024) Zaporozhets, Iryna; Musil, Félix; Kapil, Venkat; Clementi, Cecilia; Center for Theoretical Biological Physics
    The contribution of nuclear quantum effects (NQEs) to the properties of various hydrogen-bound systems, including biomolecules, is increasingly recognized. Despite the development of many acceleration techniques, the computational overhead of incorporating NQEs in complex systems is sizable, particularly at low temperatures. In this work, we leverage deep learning and multiscale coarse-graining techniques to mitigate the computational burden of path integral molecular dynamics (PIMD). In particular, we employ a machine-learned potential to accurately represent corrections to classical potentials, thereby significantly reducing the computational cost of simulating NQEs. We validate our approach using four distinct systems: Morse potential, Zundel cation, single water molecule, and bulk water. Our framework allows us to accurately compute position-dependent static properties, as demonstrated by the excellent agreement obtained between the machine-learned potential and computationally intensive PIMD calculations, even in the presence of strong NQEs. This approach opens the way to the development of transferable machine-learned potentials capable of accurately reproducing NQEs in a wide range of molecular systems.
  • Thumbnail Image
    ItemEmbargo
    Molecular coarse-graining for classical and quantum systems
    (2024-04-19) Zaporozhets, Iryna; Kolomeisky, Anatoly; Clementi, Cecilia
    Understanding the intricate molecular mechanisms underlying biological processes is crucial for tackling multiple biomedical challenges. Molecular dynamics serves as a "computational microscope", offering insights into biomolecular processes with unparalleled spatial and temporal resolution. Yet capturing these processes on biologically relevant scales poses significant computational challenges, especially when additional phenomena, such as nuclear quantum effects (NQEs), must be considered. However, many processes of interest can be described by a smaller set of collective variables instead of the intractably large number of degrees of freedom arising in atomistic simulation. The idea behind coarse-graining is to integrate out the irrelevant degrees of freedom and model the target system at a lower resolution while preserving the target properties. This thesis contributes to the development and application of coarse-grained models to increase the computational efficiency of biomolecular simulation and extend the range of molecular processes that can be investigated computationally. First, we applied a structure-based coarse-grained model combined with all-atom simulations to elucidate the helix formation mechanism following the chromophore isomerization in cyanobacteriochrome Slr1393-g3. Our findings indicate a destabilization of the helical state in the 15-Z configuration compared to the 15-E configuration, which has implications for future experimental investigations. This project also highlights the need for improved coarse-grained models. Second, the ODEM optimization framework was used to parameterize protein structure-based models using experimental data. The results suggest that incorporating many-body terms to describe nonbonded interactions is crucial to accurately reproduce the protein thermodynamics. This result underscores the importance of using neural networks' potential in approximating coarse-grained force-fields for future research. Next, a combination of coarse-graining, path integral quantum mechanics, and machine learning was used to develop potentials that incorporate NQEs into all-atom simulation at the cost of classical molecular dynamics. We developed separate models to approximate quantum dynamics and quantum statistics, which demonstrated good performance when applied to test systems. These approaches have the potential to obtain an accurate incorporation of NQEs in biomolecular simulation. Finally, we discuss how the developed approaches contribute to the bigger goal of effective and accurate methods for computational elucidation of biomolecular processes.